Human walking is characterized by pronounced arm movement, yet computer simulation models of walking usually lump the mass of the arms with the head and torso. The implications of this simplification have not been thoroughly documented in the literature. Thus, the purpose of this study was to establish the dependence of several biomechanical and energetic variables on suppressing arm swing (AS) in walking. Eight healthy adult subjects walked with and without normal AS, with speed and stride frequency/length matched between trials. Metabolic data were collected during walking on a treadmill, while kinematic and kinetic data were collected during overground walking. Gross and net energy expenditure were significantly higher during walking without AS, with the mean differences being less than 10%. Joint angles, angular velocities, and ground reaction forces were nearly identical for walking with and without AS. Most joint moments and powers were also similar between AS conditions; however, some kinetic variables (e.g., knee joint power) exhibited larger differences, primarily during the stance phase. The variable that differed most between walking with and without AS was the free vertical moment between the foot and ground. In summary, most variables differed by less than 10% and were highly correlated (r0.90) between walking with and without normal AS. Thus, researchers may be justified in using walking models without articulated arms. However, a few variables exhibited larger differences, which might be of relevance based on the specific research question being addressed.